Invasive Drosophila suzukii facilitates Drosophila melanogaster infestation and sour rot outbreaks in the vineyards

How do invasive pests affect interactions between members of pre-existing agrosystems? The invasive pest Drosophila suzukii is suspected to be involved in the aetiology of sour rot, a grapevine disease that otherwise develops following Drosophila melanogaster infestation of wounded berries. We combined field observations with laboratory assays to disentangle the relative roles of both Drosophila in disease development. We observed the emergence of numerous D. suzukii, but no D. melanogaster flies, from bunches that started showing mild sour rot symptoms days after field collection. However, bunches that already showed severe rot symptoms in the field mostly contained D. melanogaster. In the laboratory, oviposition by D. suzukii triggered sour rot development. An independent assay showed the disease increased grape attractiveness to ovipositing D. melanogaster females. Our results suggest that in invaded vineyards, D. suzukii facilitates D. melanogaster infestation and, consequently, favours sour rot outbreaks. Rather than competing with close species, the invader subsequently permits their reproduction in otherwise non-accessible resources and may cause more frequent, or more extensive, disease outbreaks.


Introduction
Drosophila suzukii (Matsumara), the spotted wing drosophila, is an invasive Asian pest that was initially detected in Europe and North America in 2008 [1][2][3]. Unlike the vast majority of Drosophila species, which oviposit in decaying fruit, 2017 The Authors. Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited.

Material and methods
All experiments and observations were conducted in a controlled climatic chamber with 14 h daylight per 24 h, 23°C and 75% RH.

Natural infestation in the field
This first experiment focused on the relationship between D. suzukii, D. melanogaster and sour rot in the field. We collected ripe grape bunches of the Meunier cultivar in Champagne vineyard (GPS coordinates 49.018, 3.982) where both D. suzukii and sour rot disease were present. We randomly sampled nine bunches without sour rot symptoms and nine bunches with severe sour rot symptoms (strong vinegar smell, typical reddish colour of the berries). Bunches were then individually placed in cages (cylinder of 25 cm diameter and 40 cm height, top covered with a 0.2 mm mesh net) where fly eggs, larvae and pupae present at the time of collection could develop, in the laboratory conditions described above. Adult flies emerging from the grapes were counted daily for two weeks and their sex and species identified. Eight days after collection, sour rot symptoms were re-evaluated. We determined three disease states of fruit degradation: -No sour rot: the healthy-looking bunch at time of field collection still had no sour rot symptoms after one week. -Post-harvest mild rot: the bunch was healthy when harvested, but showed symptoms of sour rot after one week. -Severe rot: the bunch was harvested with severe sour rot that worsened for the duration of the experiment.

Presence of sour rot agents in wild Drosophila suzukii flies
As we collected grape bunches in the field, we also collected and isolated 15 wild D. suzukii flies from the same vineyard by net sweeping between plant rows. These wild flies were used to identify whole bodyassociated bacteria by high-throughput sequencing of a DNA amplicon coding for the 16S ribosome RNA gene. We used universal primers to amplify a 251-bp portion of the V4 region of the 16S rRNA gene (16S-V4F:587 GTGCCAGCMGCCGCGGTAA; 16S-V4R: GGACTACHVGGGTWTCTAATCC) and a dual-index method to multiplex our samples [17,18]. Laboratory preparation for DNA extraction, PCRs (in duplicate), and library preparation was performed as in [18].
With the pipeline FROGS [19], 2 × 251 bp paired-end sequencing data from the Illumina MiSeq system were processed. Briefly, we trimmed 3' primers with cutadapt [20], merged paired sequences into contigs with FLASH [21], filtered sequences by length (expected value of 251b ± 10b), dereplicated sequences, removed chimaeras using the algorithm of Edgar et al. [22] implemented in VSEARCH, clustered sequences with SWARM using a local clustering threshold [23] and returned taxonomic affiliation for each OTU using NCBI Blast + on Silva SSU 119 and 123.
Filtering for false positives was performed following the strategy proposed by Galan et al. [18]. In short, we discarded positive results associated with sequence counts below two OTU-specific thresholds, which controls respectively for cross-contamination (using six negative controls for extraction and PCR) and incorrect assignment due to the generation of mixed clusters during the sequencing (using four alien positive controls of either Borrelia burgdorferi or Bartonella taylorii). Finally, flies were considered positive for a given OTU only if both technical replicates were positive after the filtering steps.

Experimental infections
To determine whether D. suzukii can trigger sour rot, we exposed single grapes to one of four different treatments: -with D. suzukii, -with D. melanogaster, -with both species, and -with no fly.
Pristine single grape berries of the Red Muscat cultivar were isolated and checked for evidence of wounds, and the stems were still strongly attached therefore preventing access of flies to the berry flesh. In order to homogenize microbial communities on experimental fruits, we bleached berries in a 2.6% sodium hypochlorite solution and rinsed them in distilled water. We then dipped each berry in one out of four 'sour rot extract' which were produced by crushing four infected bunches collected in the infested vineyard for the previous experiment. Rot extracts were diluted 10 times in distilled water to reduce microbial inoculum concentration. Even if we kept track of rot extract identity, this factor had no influence on experiment outcomes and was discarded from statistical analyses.
Individual berries were placed in a cage (cylinder of 10.5 cm diameter and 8 cm height, top covered with a 0.2 mm mesh net) with either six D. suzukii females, six D. melanogaster females, six females from each species, or no fly. Drosophila suzukii flies were collected near Montpellier, southern France, in 2013 (PL-Mu strain). Drosophila melanogaster flies were the OregonR strain, founded in 1927 in the USA. Flies had been reared on banana medium for at least 2 years (electronic supplementary materials, table S1) and had never been exposed to grapes. They were not exposed to any specific treatment before the experiment and probably had similar bacterial community [24]. There were 17 single grape berries per treatment, each forming an experimental replicate.
After fly exposure for 72 h, the number of eggs laid by each Drosophila species was counted by visual inspection where egg species can be recognized by respiratory tube morphology. We monitored berries for sour rot symptoms appearance for one week. Another grapevine disease, the grey rot, (caused by Botrytis cinerea) developed on some fruits and was recorded. This fungus has a broad host range and a different aetiology to sour rot as it does not need wounds to develop [25]. It was present in two out of the four sour rot extracts mentioned above. Grey rot development was independent from fly treatment and was discarded from statistical analyses.

Behavioural assays
Two-choice tests were conducted with D. melanogaster females to determine whether they preferred laying eggs on wounded grapes infested with D. suzukii larvae or developing sour rot. We set up two experiments: -choice between a sound grape and a grape inoculated with sour rot, and -choice between a sound grape and a grape infested with D. suzukii larvae.
Danlas cultivar grapes were sanitized by washing in a 2.6% sodium hypochlorite solution and rinsed in distilled water. Grape berries were cut to expose 1 cm 2 of flesh. We initiated sour rot disease by inoculating the wound with 500 µl of sour rot extract (see above). Drosophila suzukii infested grapes were obtained by transferring 10 D. suzukii eggs (PL-Mu strain) on exposed fruit flesh. Grapes were inspected until larvae emerged, and only grapes with at least five hatched larvae were used. Twenty-four hours after berry treatment, one D. melanogaster (OregonR strain) female was placed in a cage (cylinder of 10.5 cm diameter and 8 cm height, top covered with a 0.2 mm mesh net) with one of the two-choices tests. Females were removed the following day (ca 12 h later) and D. melanogaster eggs counted. Each treatment had 27 replicates, each involving a different D. melanogaster female.

Natural infestations in the field
We analysed separately the sums of D. suzukii and D. melanogaster adults that emerged from the grapes with generalized linear models. 'Disease state' was treated as a nominal factor, we specified a Poisson distribution and the log link function. We further used the least-square contrasts to determine differences between treatment levels.

Experimental infections
We used a nominal logistic regression to analyse the proportion of fruits with sour rot or grey rot symptoms. We first analysed the effect of the exposure treatment (nominal factor). In a second stage, using only treatments with D. suzukii females, we analysed the effect of D. suzukii egg presence (i.e. whether or not we could observe eggs) on disease development. Pairwise test using Odds ratios were used to identify significant differences in disease prevalence among treatments levels.

Behavioural assays
For each of the two-choice experiments, the number of D. melanogaster eggs deposited on each berry was analysed with a pairwise Wilcoxon's signed-rank test. All analyses were carried out within JMP Pro (SAS, 12.0.1).

Natural infestation in the field
In this first experiment, we monitored the emergence of Drosophilae from bunches collected in a vineyard in Champagne with both the presence of D. suzukii and sour rot. The number of emerging D. suzukii adults was significantly influenced by Disease state (χ 2 = 10.02; p = 0.0067; figure 1a). More D. suzukii flies emerged from bunches with Post-harvest mild rot than with No rot or Severe rot (contrast: χ 2 = 9.25; p = 0.0024).
The number of emerging D. melanogaster adults was significantly influenced by Disease state (χ 2 = 30.60; p < 0.0001; figure 1b). Drosophila melanogaster almost only emerged from bunches with Severe rot (contrast:

Vectoring of sour rot agents by wild Drosophila suzukii flies
We also collected D. suzukii individuals in the Champagne vineyard to identify their bacterial communities using high-throughput sequencing. Overall, PCRs generated a minimum of 2000 reads per sample after filtering, for an average of 13 000 reads per sample, all the samples reached the plateau (see rarefaction curves in the electronic supplementary material, figure S1). Out of the 15 flies tested, we discarded the three which were positive for Wolbachia spp. because they yielded fewer than 2000 reads for other bacterial genera (for both amplification replicates). We assumed that the small number of sequences per sample would limit the completeness of bacterial detection. In the 12 remaining samples, we considered an OTU was present in a sample when the corresponding sequence was positive. All the samples contained acetic acid bacteria (table 1) and ten contained Gluconobacter spp. and Acetobacter spp., the most effective agents of sour rot [12].

Experimental infections
We conducted trials to determine whether four exposure treatments (D. suzukii, D. melanogaster, both species or nothing) could trigger sour rot on a pristine grape berry. Exposure treatment had a significant effect on the presence of sour rot symptoms (χ 2 = 24.96; p < 0.0001; figure 2), but not on grey rot development where 15% of grapes had grey rot distributed equally among treatments (χ 2 = 0.47; p = 0.92). The two treatments with D. suzukii females (i.e. D. suzukii and Both species) induced sour rot symptoms more frequently than in the absence of D. suzukii (Odds ratios; p < 0.0039). When limiting the analysis to grapes exposed to D. suzukii females, all berries with eggs exhibited sour rot symptoms while only 15% of them did in the absence of visible eggs (χ 2 = 21,23; p < 0.0001).

Discussion
Drosophila suzukii emerged from grape bunches collected in the field more frequently when berries were healthy at the time of collection but showed mild symptoms of sour rot after one week in the laboratories (figure 1a). By contrast, numerous D. melanogaster emerged from bunches with severe sour rot symptoms in the field. Very few D. melanogaster flies emerged from bunches that appeared healthy when harvested,   whether or not the symptoms were expressed later on (figure 1b). Meta-barcode analysis revealed that most D. suzukii adults from the field carried acetic acid bacteria (table 1). Experimental infections in the laboratories showed that exposure to D. suzukii females can trigger sour rot in healthy, intact fruit. Sour rot symptoms always appeared when D. suzukii females had laid eggs in the berry (figure 2). Finally, behavioural assays showed that ovipositing D. melanogaster females preferred sour rot-infected grapes, but not berries containing D. suzukii larvae (figure 3). Our data confirm previous observations that D. suzukii is able to lay eggs in unwounded grape [4,9,10]. The high numbers of D. suzukii that emerged from bunches which developed sour rot in the laboratory (while looking healthy during collection) shows that D. suzukii and sour rot disease associate in the vineyard. Our laboratory experiments further confirmed that D. suzukii oviposition can trigger the onset of sour rot (figure 2), at least when acetic acid bacteria are present on the fruit surface as is common in the field [26]. In this assay, sour rot symptoms appeared in the presence of D. suzukii females and always occurred when D. suzukii eggs were observed (figure 2). Symptoms of sour rot when D. suzukii eggs were not observed are probably due to the difficulty of detecting deeply buried eggs.
The abundance of D. melanogaster adults that emerged from field bunches collected with rot symptoms contrasts with their absence from bunches that were healthy at the time of collection. This observation, in combination with results from experimental infections where D. melanogaster alone did not trigger sour rot symptoms on pristine grapes, is in agreement with previous publications that concluded the species cannot trigger the disease in unwounded grapes [11,12]. Before D. suzukii invaded French vineyards, wounds were due to climatic factors (for example, hail, heat shock, heavy rain leading to fruit skin shear or berry splitting) or physical damages by birds or wasps. As D. suzukii is present in the vineyard, these prior conditions seem no longer necessary as D. suzukii oviposition would suffice for disease onset.
Our behavioural assays results explain why high numbers of D. melanogaster flies emerged from bunches with severe sour rot symptoms. Ovipositing D. melanogaster females were neither attracted nor  females (figure 3a). This was not surprising as this species of fly is known to be strongly attracted to acetic acid [27].
By invading the vineyard agrosystems, D. suzukii facilitates both D. melanogaster reproduction and sour rot disease progression. Our results suggest that oviposition and larval development of D. suzukii induces sour rot disease in initially pristine grape berries, even in the absence of D. melanogaster. During early sour rot development, associated odours could attract D. melanogaster females that would also lay their eggs and colonize the fruit. This would lead to heavy infestations of D. melanogaster and large sour rot development ( figure 4). The combined influence of D. suzukii and D. melanogaster would then increase the prevalence of sour rot disease in vineyards and worsen the damage it produces.
In most years, sour rot disease in European vineyards is sporadic where notable disease outbreaks depend on weather during grape ripening, high humidity levels and low temperature favouring the disease [28]. These climatic conditions are also favourable to the reproduction of D. suzukii [29,30]. This implies future outbreaks of sour rot might be more intense and more frequent than before D. suzukii invasion as favourable factors are similar.
In this study, we show for the first time that D. suzukii participates in the onset of a host plant disease. We also highlight how tripartite interactions between flies, plants and microorganisms are affected by the introduction of an invasive species, D. suzukii. The pest, by wounding the fruit during oviposition and inducing rot onset, allows the closely related species D. melanogaster to infest grape berries that would otherwise not be accessible. The invader therefore opens a new ecological niche for the native insect and its associated microbial community. Such positive effects of invasive alien species for native species are poorly documented; most studies focus on the negative effects of invasions on communities [31,32]. There are, however, other examples of facilitative effects of invasive species. For example, the non-native meadow Lolium multiflorum facilitates the native hemipteran pest Stenotus rubrovittatus in rice crops in Japan by being a host for the pest [33]. Gandhi & Herms [34] point out that invasive species might sometimes have positive effects on entire ecosystems, especially compared to the detrimental effects eradication methods could have. D. suzukii has a large spectrum of hosts, both cultivated and wild [5]. It is thus likely that invasion by this species has led to the spread of other wild plant diseases not yet documented and with unanticipated impacts on natural ecosystems.
Ethics. Fieldwork in the experimental vineyard was authorized by CIVC. No special collecting permit or 'Animal Care Control' was required.